Multivariate State Estimation Technique (MSET) Modeling for Assured Measurement Integrity

60387 Assuring the safe, reliable, and efficient operation of a nuclear power plant requires that accurate, online measurement information be available to operators. In this project, an innovative process modeling technique will be used to detect sensor and equipment failures with higher accuracy and faster response time than previously available methods, thereby improving efficiency, reducing downtime, and lowering operating and maintenance cost for power-generating facilities. Phase I demonstrated the techniques and analytical metrics necessary to optimize a diagnostic model of a monitored process and quantify its expected performance. The analytical metrics were shown to be effective measures of model quality. Innovations included new techniques to select optimal model settings, predict in-service sensitivity and false alarm rate, and perform model self-checking. The feasibility and benefit of optimizing the model for improved detection of signal faults and equipment failures in nuclear power plant applications was demonstrated. Phase II will create a general-purpose online diagnostic software package that enables the rapid, cost effective deployment of this technology. The effort will be coordinated with nuclear power industry efforts to deploy the technology. Benefits will be quantified for nuclear plant online signal validation, thermal efficiency optimization, and equipment fault detection. Commercial Applications and Other Benefits as described by the awardee: Online signal validation should benefit any monitored process where time-critical safety or control functions depend on sensor input, or where unexpected process interruptions due to sensor and equipment failures or false alarms are unsafe or uneconomical. A large commercial market should result from improvements in the efficiency, safety, and performance of power generation, medical, aerospace, and military systems